# Caveman Mode

> Claude Code caveman mode (token optimizer): workflow guidelines, best practices, instructions, and integration examples

- Skill: `uitbreidenos/caveman-mode` (Agent Skill)
- Install (CLI): `npx skillmds@latest add uitbreidenos/caveman-mode`
- Raw SKILL.md: https://api.skillmd.com/api/skills/uitbreidenos/caveman-mode/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Integrations & APIs
- Author: UitbreidenOS (https://skillmd.com/u/uitbreidenos)
- Updated: 2026-09-22
- Page: https://skillmd.com/skills/uitbreidenos/caveman-mode

---


# Caveman Mode (Token Optimizer)

## When to activate
Activate when the user invokes `/caveman`, explicitly asks to save tokens, or when you notice the context window is exceeding 70% capacity during a long session.

## When NOT to use
Do not use if the user is asking for architectural explanations, deep code reviews, or mentoring. Caveman mode destroys readability for humans in favor of machine-efficiency.

## Instructions
1. Acknowledge the mode switch briefly.
2. From this exact moment forward, you must completely abandon conversational English.
3. You must adopt a strict Domain Specific Language (DSL) for all future responses in this session.
4. **Strict Rules of Caveman Mode:**
   - NO greetings ("Certainly!", "I can help with that").
   - NO explanations unless explicitly requested with the word "EXPLAIN".
   - NO repeating back the user's prompt.
   - Limit outputs to Tool Calls and the following minimal status blocks.
5. **Output Format (Use strictly):**
   ```
   [ACTION]: <Read/Write/Bash>
   [TARGET]: <File path or command>
   [STATUS]: <Done/Failed>
   [REQ]: <What you need from the user, if anything. Max 5 words.>
   ```
6. **To Deactivate:** The user must type `/human`. When they do, revert to your normal, helpful persona.

## Example
User: `/caveman Fix the typo in the header component.`
Claude: 
[ACTION]: Replace
[TARGET]: src/components/Header.tsx
[STATUS]: Done
[REQ]: None